The trend toward personalized management of diabetes has focused attention on the differences among available pharmacological agents in terms of mechanisms of action, efficacy, and, most important, safety. Clinicians must select from these features to develop individualized therapy regimens. In June 2013, a nine-member Diabetes Care Editors’ Expert Forum convened to review safety evidence for six major diabetes drug classes: insulin, sulfonylureas (SUs), thiazolidinediones (TZDs), glucagon-like peptide-1 receptor agonists, dipeptidyl peptidase-4 inhibitors, and sodium glucose cotransporter 2 inhibitors. This article, an outgrowth of the forum, summarizes well-delineated and theoretical safety concerns related to these drug classes, as well as the panelists’ opinions regarding their best use in patients with type 2 diabetes. All of the options appear to have reasonably wide safety margins when used appropriately. Those about which we know the most—metformin, SUs, insulin, and perhaps now also TZDs—are efficacious in most patients and can be placed into a basic initial algorithm. However, these agents leave some clinical needs unmet. Selecting next steps is a more formidable process involving newer agents that are understood less well and for which there are unresolved questions regarding risk versus benefit in certain populations. Choosing a specific agent is not as important as implementing some form of early intervention and advancing rapidly to some form of combination therapy as needed. When all options are relatively safe given the benefits they confer, therapeutic decision making must rely on a personalized approach, taking into account patients’ clinical circumstances, phenotype, pathophysiological defects, preferences, abilities, and costs.